Abstract: This paper introduces the significance of UPS energy saving and common green energy-saving technologies for UPS. It mainly covers energy-saving technologies related to UPS and batteries, including UPS power supply planning, UPS energy-saving technologies, harmonic mitigation solutions, shared battery pack solutions, intelligent battery management, and intelligent power distribution. Keywords: UPS; energy saving; high efficiency; low harmonics; shared battery pack; intelligent battery management solution I. Overview With the rapid economic development and enterprises' increasing understanding of the Internet, data centers have flourished in recent years. However, this has led to increasingly exorbitant electricity costs. As shown in the figure below, electrical, power supply, and cooling costs account for more than half of the investment in a data center, with electrical costs alone reaching 26%. This high energy consumption keeps the overall operating cost of data centers high, leading to the awkward situation of "being able to build but not being able to afford to use." Recent forecasts from IDC and Gartner indicate that by 2010, enterprises will spend more on electricity costs annually than they invest in hardware equipment that year. Currently, international energy supplies are becoming increasingly tight, with rising prices for various raw materials and electricity, which will exacerbate electricity costs for data centers. How to reduce data center operating costs has become a key concern for CIOs of various enterprises. Where does all the electricity go? It's clearly consumed by the various IT equipment in the data center—servers, terminals, network equipment, and cooling systems. Energy conservation can be approached from top to bottom, focusing on several key aspects. First, energy conservation in the data center environment, including cooling and power supply; second, energy conservation in IT hardware, reducing the energy consumption of IT equipment; and finally, energy conservation in the integrated circuits within IT equipment, such as CPUs. UPS is the most crucial link in the AC power supply chain; almost all IT equipment in a data center requires UPS power, and the total installed capacity of UPS in large data centers has reached the megavolt-ampere level. Improving operational energy efficiency is imperative. Currently, UPS energy conservation must be addressed comprehensively from the aspects of solution design, UPS, batteries, and power distribution. II. Flexible Planning On-demand expansion in data centers typically involves considering the needs of the next few years before implementation. However, UPS systems are often installed all at once, with several high-power UPS units running in parallel. As a result, the initial load is only 10-20%, and the equipment becomes obsolete before reaching the planned load. This not only wastes investment but also prevents the UPS from operating at its highest efficiency, leading to wasted energy. From a UPS system perspective, avoiding this situation should include: a) Power Supply Scheme Design: Currently, there are two main UPS power supply schemes: distributed power supply and centralized power supply. Distributed power supply involves one UPS powering one or more devices, with the entire data center consisting of many such systems. The advantage of distributed power supply is that it disperses risk, preventing large-scale power outages caused by a single UPS failure. The disadvantage is that the UPS units are distributed, making management difficult and cabling planning challenging. The other option is a centralized power supply scheme, where a single high-power UPS system directly powers all loads in the data center. The advantages of centralized power supply are ease of planning, management, and maintenance. The disadvantage is that if the UPS system malfunctions, it can easily cause a large-scale power outage. This disadvantage can be avoided by using various parallel architectures. Therefore, the two solutions mentioned above each have their advantages and disadvantages. Currently, data centers generally adopt centralized power supply solutions, which also centralize the power supply risks. Since there are limitations on the number of UPS units that can be connected in parallel, and when the number of UPS units connected in parallel exceeds four, its reliability is not much higher than that of a single-unit system. When the total installed capacity of UPS units in the data center exceeds a certain limit, it is recommended to plan the data center into several areas for power supply in several phases. When planning, you can refer to the following: the capacity of a single unit should not exceed 400KVA, and the number of units connected in parallel should not exceed three. b) UPS online parallel expansion function: The UPS capacity planning of the data center can also adopt a gradual expansion scheme according to the load capacity requirements at different times, making the investment scheme more economical, while also allowing the UPS to operate at its optimal power point. Currently, medium and high power UPS units all have redundant parallel operation capabilities, which not only improves system reliability but also provides conditions for data center expansion. As long as sufficient circuit breakers are reserved in the front and rear distribution boxes of the UPS during the planning stage, and corresponding space is planned in the computer room, the UPS parallel expansion function can be realized. The key is the handling of the parallel process. Most brands require modification of the UPS circuit when paralleling, which inevitably requires the UPS to operate in maintenance bypass mode. The UPS is directly powered by the mains power. If the mains power fluctuates greatly or even fails, it will cause a large-scale paralysis of the system. Therefore, parallel expansion must have online parallel function. That is, when paralleling the UPS, you only need to modify the software of the new UPS to be consistent with the original UPS system. Without shutting down the original UPS system, the new UPS can be directly connected to the original system. The UPS operates in online mode before and after the expansion, avoiding the high-risk operation of switching to bypass power supply. c) Using modular UPS to achieve gradual expansion Currently, modular UPS has begun to be used in China. The main features of modular UPS include: expandability, short mean time to repair (MTTR), and economical N+X redundancy. Taking Delta's C-series UPS as an example, each module is 20KVA, and the entire system can be expanded to a maximum of 160KVA. Expansion can be gradually implemented based on the actual capacity requirements of the data center, as long as the power distribution capacity is planned in the initial stages of data center operation. Achieving N+X redundancy is also cost-effective. For example, to achieve N+1 redundancy with a 60KVA UPS, a traditional solution requires adding a 60KVA UPS, while with a modular UPS, only a single 20KVA module needs to be added, saving a significant amount of capital. III. Improving UPS Energy Efficiency and Optimizing Load Efficiency Curves Currently, UPS systems are all online double-conversion architectures, and both the rectifier and inverter exhibit power losses during operation. Taking a 400KVA UPS as an example, with each kilowatt-hour costing 0.95 yuan, a 1% increase in UPS efficiency results in an annual electricity cost saving of (400KVA × 0.8) × 0.01 × 24 × 365 × 0.95 = 26630.4 yuan. Therefore, improving the energy efficiency of UPS systems can save a data center a significant amount of electricity costs. Improving UPS efficiency is the most direct way to reduce overall data center energy consumption. Therefore, when purchasing UPS systems, prioritize those with higher efficiency. Of course, high UPS efficiency requires more than just high full-load efficiency; it also necessitates a high efficiency curve, especially in 1+1 parallel systems. According to system planning, each UPS's capacity should not exceed 50%. If the efficiency is below 90%, even if the full-load efficiency reaches over 95%, it's meaningless. Therefore, UPS systems must employ measures to optimize the efficiency curve, ensuring high efficiency even at lower loads. Taking Delta's C-series 20KVA UPS as an example, its full-load power is 20KVA/18KW. As shown in the graph, its efficiency is already above 90% when the load is below 2KW, and it can achieve a high overall efficiency of 95% from 6KW to 18KW. Besides improving the UPS's own efficiency, some of its functions can also be utilized, such as the ECO economic operation mode. The principle behind this function is that under favorable mains power conditions, the UPS is directly powered by the static bypass. In this state, the inverter is in standby mode, operating normally but not outputting energy. If the mains power fails, the UPS immediately switches to inverter power supply, with a switching time typically within 1ms. See the diagram below for details; blue represents the input current waveform, and yellow represents the output voltage waveform. Because the inverter is in standby mode, its own losses are minimal, resulting in an overall UPS efficiency of over 97%, saving more than 3% of power compared to normal mode. The following conditions must be met to use ECO mode: a) The static bypass must use two sets of highly reliable SCR transistors. A combination of contactors and SCR transistors is not permitted, as contactors will arc when engaged, typically failing after several hundred cycles. SCR transistors do not have this problem and also shorten the switching time. b) It is recommended to use this mode in environments with good power conditions, such as primary power supply units. IV. Reducing Input Current Harmonics and Improving Power Factor The fundamental cause of harmonic generation is that power lines have a certain impedance, equivalent to a passive network composed of resistors, inductors, and capacitors. Non-sinusoidal currents generated by non-linear loads cause distortions in current and voltage in the circuit, known as harmonics. The hazards of harmonics include: causing additional losses and heat generation in electrical components (such as capacitors, transformers, motors, etc.); increased temperature and efficiency of electrical components, accelerated insulation aging, and reduced service life; interference with normal equipment operation; increased reactive power factor, reducing the active power capacity of power equipment (such as transformers, cables, distribution equipment); low power supply efficiency; resonance, especially severe when using generators; circuit breaker tripping, fuse blowing, and unexplained equipment damage. A UPS is a non-linear load for the power grid, generating a large number of harmonics during operation. Taking a UPS with a 6-pulse rectifier as an example, its input power factor is generally around 0.75, and harmonics are greater than 30%. The main methods to reduce UPS operating harmonics are: a) Using a 12-pulse rectifier. The principle is to add a phase-shifting transformer and a 6-pulse rectifier to the input of the original 6-pulse rectifier. This technology can reduce harmonics to about 10%. The advantages are its simplicity and significant harmonic improvement; the disadvantages are limited improvement in power factor and a slightly higher price. b) Using a passive filter. This uses an LC filter circuit to filter out harmonics generated by the UPS and compensate for the power factor. The advantages are its simplicity and low cost; the disadvantages are that it can only compensate for specific orders of harmonics and is greatly affected by load impedance, making it unsuitable for the entire power range. c) Using an active filter. The principle is to use controllable power semiconductor devices to inject a current into the grid with the same amplitude but opposite phase to the harmonic source current, making the total harmonic current of the power supply zero, thus achieving real-time harmonic current compensation. The advantages are that it can compensate for multiple orders of harmonics and is not affected by the load impedance. The disadvantage is the higher purchase cost. d) Using a high-frequency IGBT rectifier and a PFC power factor correction circuit to design the rectifier. The principle is to use high-frequency PWM to control the IGBT conduction, segment the input voltage waveform, make the input current waveform as close to a sine wave as possible, and compensate for the phase difference between the input voltage and current. The advantages are light size, low price, and good performance. The disadvantages are complex technical structure, difficulty in maintenance, and current capacity limitations due to power device limitations. The performance and investment comparison of the above technologies are as follows; a suitable solution can be selected based on actual needs . V. Battery Management and Power Distribution Technology UPS systems are equipped with batteries, and users often invest a large proportion of the total UPS system investment in battery packs, even exceeding the investment in the UPS itself. However, the lifespan of batteries is significantly shorter than that of the UPS main unit. Since the main materials of batteries are heavy metals like lead, sulfuric acid, and non-degradable plastics, they all cause serious environmental pollution. Therefore, reducing the number of batteries used and extending their cycle life not only saves direct and indirect battery investment but also reduces the pollution of the entire computer room to the social environment. Therefore, UPS systems can achieve battery energy saving through the following technologies: a) Parallel sharing of battery packs. The principle of shared battery packs is to use special rectifier control and fault isolation technology to synchronize the rectification and share the current on the buses of two or more UPS units in a parallel system. This allows the buses of each UPS unit in the system to be directly connected in parallel. Then, batteries that meet the system's backup time requirements are connected in parallel to the parallel bus system, achieving battery sharing and reducing battery investment. Taking a 1+1 UPS as an example, in a traditional UPS solution, the system has a backup time of 1 hour. Considering that if one UPS fails, the battery of UPS2 cannot be used by UPS1, UPS1 and UPS2 must each be equipped with a 1-hour battery pack to ensure that the system can still provide backup for 1 hour after a power outage. With the shared battery pack solution, because the batteries in the system can still provide energy to UPS2 after UPS1 fails, the entire system only needs to be equipped with one 1-hour battery pack. This not only saves direct battery investment but also saves investment in computer room space, load-bearing capacity, and air conditioning, and also reduces environmental pollution. b) Intelligent battery management technology. Many factors affect battery life, mainly including temperature, charging, discharging, and cycle count. If the above factors can be properly addressed, battery life can be significantly extended, battery replacement cycles can be increased, and battery investment can be saved. UPS intelligent battery management technology mainly includes battery equalization and float charging management (equalization and float charging control), charging temperature compensation, and intelligent discharge cutoff voltage control. In addition, it should also have periodic automatic battery detection and battery leakage detection functions. Furthermore, a UPS with a wider input voltage range can be selected to reduce the number of battery discharge cycles. Through these technologies, battery life can be significantly extended by 2-3 years. c) Intelligent UPS power distribution management technology. The principle is to detect the UPS battery voltage or equipment power supply time to achieve multiple power-down protection functions for different levels of loads in the computer room, reducing battery investment and improving battery utilization. Intelligent UPS power distribution management technology mainly has two solutions: software implementation and hardware implementation. Taking Delta UPS as an example, its software implementation involves installing the Delta Shutdown Agent on the load server in the UPS monitoring network. When the mains power is abnormal and the battery voltage or timing conditions are met, the shutdown agent automatically saves the system program and then shuts down the server. The hardware-based UPS output is configured with an intelligent power distribution panel. This panel uses a PLC to detect UPS battery voltage or timing requirements. When these conditions are met, the intelligent power distribution panel shuts down a specific output according to pre-set timeouts. This solution has already been implemented in the UPS power supply systems of several subway lines in China. VI. Conclusion Data center energy conservation must be addressed from top to bottom, or from infrastructure to core equipment. The UPS is the core of the entire AC power supply chain. Effective UPS energy conservation can not only save significant equipment investment and maintenance costs but also substantially reduce subsequent operating costs. Of course, UPS energy conservation requires joint efforts from users and manufacturers. Currently, UPS manufacturers have launched their own products and solutions; customers only need to tailor their planning accordingly.